Research On Some Key Issues In Context-Aware Computing

Pervasive computing, whose objective is to build an open and ubiquitous environment to provide users with intelligent and personalized information and services, brings us from the device-centered computing model to user-centered computing model. Context-aware computing is an important area of pervasive computing. This dissertation studies the architecture of context-aware system, context modeling, context acquisition and the application of context. The main contributions are as follows:Current research on context aware computing is focused mainly on small scale environment, while research on large scale context aware computing is a lack of support. This dissertation analyzes the feasibility of building a context aware service platform on network management system and the related problems to be solved are pointed out. Then we propose a context aware service platform based on network management system (NM-CASP). Also we introduce the architecture, system design and service process of NM-CASP.Context modeling is the basis and precondition of context-aware computing. Ontology-based context modeling technology is the focus of current research. There are still some deficiencies in the current context model based on ontology, especially in the expression of temporal and topological relations of space. This dissertation proposes a modeling process for context modeling. Then the capability to express temporal relations for context model is enhanced with Allen's Interval Algebra and Interval Gap theory. Moreover, Extended RCC (ERCC) theory is proposed which can express the topological relations for space with Empty Hole and Exclave. Finally, ontology-based location context is extended with ERCC.Location based service is the important application of context aware computing, while location context is the important context. Acquisition of indoor location context still needs to improve. After introducing WLAN-based positioning algorithms and comparing their respective advantages and disadvantages, a fingerprint matching positioning algorithms based on KSVM is proposed. Experimental results show that the proposed algorithm can improve positoning accuracy compared with KNN and SVM. And also the proposed algorithm is faster than SVM.The ultimate purpose of context-aware computing is to apply context to applications. This dissertation presents the applications of our context-aware research achievement in "Intelligent Integrated Service Platform", "General Service Platform" of China's Next Generation Internet, VHE and plan-based context aware tourist system. Acorrding to the project requirement several context-based adaption functions were provided, which included session adaption and terminal adaption. These adaption functions were implemented in our prototype.